{"id":"W4220991771","doi":"10.1145/3506712","title":"Machine Learning and Data Cleaning: Which Serves the Other?","year":2022,"lang":"en","type":"article","venue":"Journal of Data and Information Quality","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":51,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Data quality; Software deployment; Data science; Data integration; Quality (philosophy); Data mining; Software engineering; Engineering","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.07152136,0.001742522,0.005239761,0.006614832,0.005035177,0.01717764,0.005089624,0.01266671,0.007506047],"category_scores_gemma":[0.1120041,0.001193143,0.002486946,0.01043934,0.02308006,0.04361614,0.01189791,0.01741279,0.005002258],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005777326,"about_ca_system_score_gemma":0.01080583,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007692141,"about_ca_topic_score_gemma":0.005181109,"domain_scores_codex":[0.9439567,0.03569829,0.002692401,0.005343246,0.01026506,0.002044325],"domain_scores_gemma":[0.8591707,0.08854808,0.007952981,0.01292209,0.02558225,0.005823916],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002933572,0.000285069,0.008499325,0.005815925,0.0006314433,0.0003063944,0.003838439,0.001231206,0.0006543815,0.2313642,0.151699,0.5953813],"study_design_scores_gemma":[0.0001029928,0.0002075165,0.004055777,0.008103324,0.0002461692,0.001064917,0.00860488,0.005275103,0.001073999,0.5980473,0.3729418,0.000276191],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.001690619,0.2046066,0.06761272,0.7113822,0.007269031,0.00008128655,0.0001302913,0.0003295019,0.006897769],"genre_scores_gemma":[0.1398869,0.4284501,0.1501798,0.2071728,0.0614489,0.0005220362,0.0004801283,0.0009140194,0.01094546],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.07152136,"threshold_uncertainty_score":0.3782457,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.3220743629101961,"score_gpt":0.472009332225758,"score_spread":0.1499349693155619,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}